GESIS Study number ZA7576 ICPSR Study Number [not yet assigned]

EUROBAROMETER 91.5

Carried out by Kantar Public Brussels on request of the EUROPEAN COMMISSION, Directorate-General Communication (COMM.A.3.: Media monitoring and Eurobarometer) and the EUROPEAN PARLIAMENT, Directorate General Communication, between June 7 and July 1, 2019.

Archive pre-release, dataset version 1.0.0 as of December 09, 2019

This dataset edition has not yet passed the complete archive processing and documentation, such as complete variable documentation for online data browsing and analysis in ZACAT (https://zacat.gesis.org). Basic archival checks have been performed; technical variables, protocol variables and variable labels have been adapted to archive standards for the EUROBAROMETER series.

Proposed citation

European Commission and European Parliament, Brussels: Eurobarometer 91.5, June-July 2019. Kantar Public, Brussels [Producer]; GESIS, Cologne [Publisher]: ZA7576, dataset version 1.0.0, doi:10.4232/1.13393

Survey instruments

For question wording, answer category text, interviewer instructions or question routing, please see the BASIC BILINGUAL QUESTIONNAIRE and the respective national field questionnaires. The basic questionnaire also includes a technical description of this EUROBAROMETER survey.

The national field questionnaires in the different language versions are available through the GESIS Eurobarometer web pages (STUDY PROFILES), the ZACAT online data catalogue, or the Eurobarometer study descriptions in the data holding catalogue (DBK), as soon as all embargos have been lifted, if applicable.

Weighting information

In general the Standard and Special Eurobarometer datasets provide for two types of weighting, a post-stratification weighting and a population size weighting.

For each sample (country or lower level region) through the post-stratification weighting procedure a comparison is carried out between the sample composition and a proper universe description. The universe description is made available by the National Survey Research Institutes and/or by EUROSTAT. On this basis a national weighting procedure, using marginal and intercellular weighting, is applied. As such in all countries, minimum sex, age, region NUTS II (basic regions as defined by the EUROSTAT nomenclature of territorial units for statistics), and size of locality are introduced in the iteration procedure. This post-stratification weighting is also referred to as redressement or non-response weighting. A design weight which would adjust for unequal selection probabilities (depending on the household size) is not made available.

The population size weighting factor corrects for the fact that most samples are of almost identical size, no matter how large or small the populations are from which they were drawn. These weights ensure that each country as well as each lower level sample (East and West Germany) are represented in proportion to its population size within different country/sample groupings, or according to the historical states of European unification (e.g. founder members, new members, Euro zone) in the case of the EUROPEAN WEIGHTS, or within the whole country in the case of Germany (WEIGHT SPECIAL GERMANY).

The population size weights all include the post-stratification weighting factors. The EUROPEAN WEIGHTs adjust each sample in proportion to its share in the total population aged 15 and over of the European Union (formerly European Community) or in accordance with its historical compositions. These adjustments are based on population figures published by EUROSTAT in the Regional Statistics Yearbook. Between Eurobarometer 33 and 54.1 adjustments to the predefined standard sample size is taken into account. In general all samples which do not belong to the respective group of countries/samples under consideration are excluded from calculation.

The application of post-stratification weights is recommended for descriptive (univariate) analysis. Meaningful descriptive results for groups of countries or for countries with separate samples (Germany) require population size weighting. Official Eurobarometer reports are always based on weighted data.

The “overall” WEIGHT EXTRA extrapolates the actual universe (population aged 15 and over) for each country or sample. This weight variable integrates all other available weights, but does not reproduce the number of cases in the dataset, but the respective actual population size.

The following description of available standard weights in general also applies to the corresponding election versions of each weight (W31 to W79, w101 to w149, WEXMEN and WEXMEN2) which are intended for analysis of persons of voting age in the respective country’s European resp. national elections. These weights are intended for use with questions in the European Election module QG, and they include further adjustments designed to adapt the samples to the actual aggregate participation patterns observed during the respective elections.

The following weights are provided for Eurobarometer 91.5:

W1/W31/W101 (WEIGHT RESULT FROM TARGET) reproduces the real number of cases for each country. East and West German samples are weighted separately. This weight in its function corresponds to former NATION WEIGHT II (until EUROBAROMETER 31).

W3 (WEIGHT GERMANY) adjusts the East and the West German samples to their respective proportions in the united Germany. All other samples are excluded. This weight should be used whenever the united Germany is to be analysed as a whole.

W22/W52/W122 (WEIGHT EU27) includes all 25 member countries after the 2004 enlargement, and the new members as of 2007 Romania and Bulgaria; W24/W54/W124 (WEIGHT EU NMS 12) refers to the group of the twelve new members as of 2004/2007, all other samples are excluded from calculation.

W14/W44/W114 (WEIGHT EU25) refers to the EU member countries as of the 2004 enlargement; W13/W43/W113 (WEIGHT EU NMS 10) to the group of the ten new members at times. In both cases all other samples are excluded from calculation.

W11/W41/W111 (WEIGHT EU15) refers to the EU members after the 1995 enlargement. All other samples are excluded from calculation; W10/W40/W110 (WEIGHT EU NMS 3) separates the three new members at times.

W8/W38/W108 (WEIGHT EU12) excludes Austria, Sweden and Finland from the EU 15 group, and the latter two include East Germany; W9/W39/W109 (WEIGHT EU12+) includes East Germany; W7/W37/W107 (WEIGHT EU10) in addition excludes Spain and Portugal (enlargement 1986), and the latter two include East Germany; W6/W36/W106 (WEIGHT EU9+) also excludes Greece (enlargement 1981), and the latter two include East Germany.

W5/W35/W105 (WEIGHT EU6) refers to the six EC founder members: France, Belgium, the Netherlands, West Germany, Italy, and Luxembourg and the latter two include East Germany. All other samples are excluded from calculation.

W23/W53/W123 (EU28) refers to the EU 28 countries (EU27 plus Croatia; membership as of July 2013); W94/W74/W144 (WEIGHT EU NMS 13) groups the 13 new member countries for the enlargement period 2004 to 2013.

W29/W59/W129 (WEIGHT EURO ZONE 15) separates the 15 countries which introduced the EURO as of 01/01/2002 (Belgium, Germany, Ireland, Spain, France, Italy, Luxembourg, the Netherlands, Austria, Portugal, Finland and Greece), 01/01/2007 (Slovenia) and 01/01/2008 (Cyprus, Malta); W30/W60/W130 (WEIGHT NON EURO ZONE 15) refers to the rest of the EU 27 member countries which so far did not introduce the common currency, grouped as a whole. In each case all other samples are excluded from calculation.

W81/W61/W131 (WEIGHT EURO ZONE 16) separates the 16 countries which introduced the EURO as of 01/01/2002 (Belgium, Germany, Ireland, Spain, France, Italy, Luxembourg, the Netherlands, Austria, Portugal, Finland and Greece), 01/01/2007 (Slovenia), 01/01/2008 (Cyprus, Malta) and 01/01/2009 (Slovakia); W82/W62/W132 (WEIGHT NON EURO ZONE 16) refers to the rest of the EU 27

member countries which so far did not introduce the common currency, grouped as a whole. In each case all other samples are excluded from calculation.

W89/W69/W139 (WEIGHT EURO ZONE 17) adds ESTONIA to the group of EURO zone countries as of January 2011. W90/W70/W140 (WEIGHT NON EURO ZONE 17 EU27) refers to the group of EU27 countries which as of January 2011 did not introduce the common currency.

W98/W78/W148 (WEIGHT EURO ZONE 18) adds LATVIA to the group of EURO zone countries as of January 2014. W83/W63/W133 (WEIGHT NON-EURO ZONE 18 EU27) references the EU27 countries, W99/W79/W149 (WEIGHT NON-EURO ZONE 18 EU28) the EU28 countries which as of January 2014 did not introduce the common currency. In each case all other samples are excluded from calculation. W84/W64/W134 (WEIGHT NON EURO ZONE 18 NMS 12) refers to the non-EURO zone 18 countries among the 12 new members as of 2004/2007. In each case all other samples are excluded from calculation.

W85/W65/W135 (WEIGHT EURO ZONE 19) adds LITHUANIA to the group of EURO zone countries as of January 2015. W86/W66/W136 (WEIGHT NON-EURO ZONE 19 EU28) references the EU28 countries, which as of January 2015 did not introduce the common currency. W87/W67/W137 refers to the EU 28 countries minus the United Kingdom.

W95 (WEIGHT EU CC4) refers to the group of all candidate countries at the time of the fieldwork, i.e. Turkey, Macedonia, Montenegro and Serbia. In each case all other samples are excluded from calculation.

W97 (WEIGHT EU CC BALKANS) separates the group of Balkan Candidate Countries Macedonia, Montenegro, Serbia and Albania. In each case all other samples are excluded from calculation.

W92 (WEIGHT TOTAL) refers to the total of all countries / samples as a whole, i.e. EU28 plus the five Candidate Countries (W100) and the Turkish Cypriot Community (Northern Cyprus).

W100 (WEIGHT EU CC5) refers to the Candidate Countries as of March 2015, i.e. Turkey, Macedonia, Montenegro, Serbia and Albania, but now excluding Iceland. All other samples are excluded from calculation.

WEX/WEXMEN/WEXMEN2 extrapolates the actual universe (population aged 15 or more) for each country (all samples), i.e. this weight variable integrates all other available weights, but does not reproduce the number of cases in the dataset.

Remarks about this dataset version

• Modules QA to QE in the standard Eurobarometer context and largely replicating questions asked in the context of Eurobarometer 90.3 (ZA7489) and Eurobarometer 89.1 (ZA6963).

• Question modules QF and QG “Post-election survey 2019” were implemented on behalf of and financed by the European Parliament. They partly replicate questions asked in the context of Eurobarometer 91.1 [ZA7561],

Eurobarometer 90.1 [ZA7484], Eurobarometer 87.3 [ZA6863], Eurobarometer 88.3 [ZA6928] and Eurobarometer 87.3 [ZA6863].

• Question module QH “EU citizens and development cooperation” partly replicates questions formerly asked in the framework of Eurobarometer 89.3 (ZA7483).

• Region NUTS 2 categories for PORTUGAL have been edited and adapted to changes as of 2002 (NUTS version 2003) which are reflected in Eurobarometer starting with wave 62 (confirmed by TNS). Region NUTS 1 categories for GREECE have been edited correcting for the non-coverage of Ionian and Aegean Islands starting with Eurobarometer 62 (confirmed by TNS).

• Starting with Eurobarometer 89.2 no separate sample is drawn for Northern Ireland, which is now covered by the sample for the United Kingdom. Accordingly, the formerly available weight variable w4 “WEIGHT UNITED KINGDOM” is obsolete and no longer included in the dataset.

• The fieldwork dates implied by the values in protocol variable p1 (date of interview) differ slightly from those indicated in the technical specifications provided by Kantar for several countries.

• Please note that variables gen1 to gen6 are based on age instead of year of birth, which may entail a slightly inaccurate categorization for some respondents.

• Data for protocol variables p8 (postal code), p9 (sample point number), and p10 (interviewer number) have not been made available.

• Extrapolated population figures as calculated by WEX differ from the population 15+ figures for Turkey indicated in the technical specifications provided by Kantar Public.

• LANGUAGE OF INTERVIEW in North Macedonia (p13mk) has been corrected according to plausibility consideration and checked against regions (p7mk) with Albanian resp. Macedonian speaking majorities.

• SIZE OF COMMUNITY (p6mk) is not available for North Macedonia.

• Variable qg1_orig ("EUROPEAN PARLIAMENT ELECTION 2019 - VOTED INCL DK/REFUSAL (APPOINTED BY THE ARCHIVE)" was appointed by the archive. It retains respondents as "DK/Refusal" that were collapsed into the category "Not voted" in the source data version of qg1 delivered to the archive for reasons of calculating the election weights. This approach has been approved by Kantar.

• Questions qg2 and qg16 refer to country-specific political parties that can be identified via the corresponding country questionnaires.

• Please note that 13 respondents are coded 2 (“Student”) in variable d15a (“OCCUPATION OF RESPONDENT”) and at the same time indicate an age when stopped full-time education in variable d8 (“AGE EDUCATION”).

According to the interviewer instruction these cases should have been coded to the answer category “Still studying” in variable d8 (“AGE EDUCATION”).

• Please note that three respondents give substantial answers to qd5 (“EU FEELING ISSUES”) and at the same time are coded 1 (“Mentioned”) in qd5.15 (“EU FEELING ISSUES: NO SUCH FEELING (SPONT)”).

• Please note that 127 respondents from the Turkish Cypriot Community are coded 1 (Mentioned) in q1.29 (“NATIONALITY: TURKEY”) even though q1.29 should only have been fielded in Turkey. Moreover, two cases from Slovenia are coded 1 (Mentioned) in q1.29 (“NATIONALITY: TURKEY”) (uniqid is 420000050) resp. q1.33 (“NATIONALITY: SERBIA”) (uniqid is 420000128) even though q1.29 and q1.33 should only have been fielded in Turkey or Serbia. Both cases do not mention any other nationality. Therefore the countries excluded by the corresponding filter statement countries are respectively coded 0 (“Not mentioned”).

• One case from Belgium is coded 1 (Mentioned) in q1.29 (“NATIONALITY: TURKEY”) (uniqid is 10000031) even though q1.29 should only have been fielded in Turkey. This case does not mention any other nationality. Therefore the countries excluded by the corresponding filter statement countries are coded 0 (“Not mentioned”).

• Two cases from the Netherlands are coded 1 (Mentioned) in q1.29 (“NATIONALITY: TURKEY”) (uniqid is 110000352) resp. q1.34 (“NATIONALITY: ALBANIA”) (uniqid is 110000306) even though q1.29 and q1.34 should only have been fielded in Turkey or Albania. Additionally, both cases are coded 1 (Mentioned) in q1.10 (“NATIONALITY: NETHERLANDS”). Therefore the countries excluded by the corresponding filter statement countries are coded 0 (“Not mentioned”).

• Checks for duplicate case identification numbers and cases have been performed: Some country specific case ids for DE-W and DE-E and all case ids for CY-TCC were missing (variable “caseid” was coded blank = system missing). However, the dataset does include an original unique serial id across all countries (variable “serialid”) appointed by Kantar. The dataset has been checked for duplicate cases on the base of 64 selected non-technical (substantial) variables without evidence for any duplicate.

GESIS – Leibniz Institute for the Social Sciences Data Archive for the Social Sciences (DAS)

Cologne, Germany December 09, 2019

Archive website: https://www.gesis.org/eurobarometer-data-service/home/

Archive contact: [email protected]

European Commission DG Communication: http://ec.europa.eu/COMMFrontOffice/PublicOpinion

European Parliament DG Communication: https://www.europarl.europa.eu/at-your-service/de/be-heard/eurobarometer/post- election-survey-2019-first-results